It works, but not the way books and gurus explain it. The technical levels (entries, stops, targets) should be symmetrical and not pattern or OHLC bar based. Real TA is 100x harder to learn, as it’s heavily grounded within a quantitative framework. You need other tools, not just charts and indicators.
I wouldn't call it a joke. They are useful in trading, especially when you are looking for relationships between price/volume indices and market indicators.
I may pass by technical but understanding why price is moving the way it does is important - price is not random. When a bar stops going higher it is because sell orders kept it down. When the bar stops going lower it is because it hit buy orders. When price crosses under or over the sells or buys it creates imbalance. This sets direction and a reaction back to that level.
I have not come across a successful TA trader in my long voyage in trading. It's been all gurus and "educators" pushing it. Not to say there aren't any profitable practitioners as you can't prove a negative. But the burden of proof - and we should all easily agree on this - is on the people who claims that TA works. As an aside, there is absolutely nothing "technical" about drawing lines on charts or moving average calculations (and all their various derivatives). The name itself - technical analysis - is one of the great misnomers.
TA is a joke for those who are in the prediction business. Those who use it as a framework to structure their bet understand what it takes to be profitable. If you don’t get some frequencies and or magnitudes outputs out of your TA then it’s probably useless TA. As I wrote here : No Structure -> No Value -> Gambling.